How to Read Your Monthly Marketing Report (and Spot the Bits That Are Hiding Something)
Most business owners do the same thing on a monthly reporting call. They nod at the charts, note that the arrows are mostly green, ask one polite question, and hang up with no clearer idea of whether the money is working than they had before.
That is not a competence problem. A lot of marketing reports are built to survive scrutiny rather than invite it: front-loaded with the metrics that move easily, structured so the uncomfortable numbers arrive on slide fourteen when everyone has stopped reading, and written in a vocabulary that makes asking a basic question feel naive.
This guide takes the vocabulary away as an excuse. It covers the order a good report should follow, what each metric actually proves, why four systems will give you four different answers about the same month, the eight questions that make a reporting call useful, and the specific red flags that mean you are being shown activity instead of results.
It sits under our performance marketing guide for Singapore, and it is the post that most directly describes how we run our own performance marketing reporting.
The one question a report exists to answer
Every marketing report is trying to answer some version of: should we keep doing this, change it, or stop?
That is a decision question, and decision questions need three inputs: what happened, why it happened, and what we are going to do about it. Anything in a report that does not serve one of those three is decoration. Reach is decoration. A word cloud is decoration. A chart of impressions by day of week is decoration wearing a lab coat.
The test is simple and slightly brutal: read a page and ask what decision it would change. If the honest answer is “none”, the page is there to fill space, or to occupy the part of the meeting where you might otherwise have asked something difficult.
The order a good report follows
Order matters more than content. The same five facts arranged two ways produce two entirely different meetings.
Note item four on the right. A report with no failures in it is not a report about a good month; it is a report that has been edited. Real marketing months contain a test that flopped, an audience that stopped responding, a landing page that underperformed. If those never appear, either nothing is being tested or the results are being filtered.
The metric decoder
Here is what each common metric actually proves — and, more usefully, how each one is made to look good without the business improving.
| Metric | What it really means | What it proves | How it gets gamed |
|---|---|---|---|
| Impressions / reach | How many times the ad was served, and how many people saw it | The media was bought. Nothing more. | Shift budget to cheap placements or broad audiences. Reach can double while enquiries fall. |
| Clicks / CTR | How many people tapped, and what share of viewers that was | The creative and targeting are not broken | Curiosity headlines and vague creative lift CTR and lower lead quality at the same time. |
| CPC / CPM | Cost per click, or per thousand impressions | How expensive the auction was | Cheaper traffic is easy to buy and usually worth less. A falling CPC alone is not good news. |
| Sessions / traffic | Visits to the site | People arrived | “Traffic up 40%” is the most common headline in weak reports. Ask which pages, from where, and what they did. |
| Conversions | Whatever the platform was told to count | Only as much as the definition allows | Counting page views, scroll depth or a tap on a WhatsApp button as a “conversion”. Ask which event fires. |
| Cost per lead | Spend divided by leads | Efficiency — if “lead” is defined tightly | Loosen the definition and CPL falls instantly. This is the most gameable number in the report. |
| ROAS | Attributed revenue divided by ad spend | Media efficiency in the platform’s own accounting | Retargeting-heavy budgets inflate it. Revenue passed with GST included inflates it further. |
| Qualified enquiries | Leads that met a written standard | The closest thing to a real number in the report | Hard to game — which is exactly why it is so often absent. |
If you take one thing from this table: ask for the definition of “lead” and “conversion” in writing, once, and hold every subsequent report to it. Most reporting disputes are definition disputes wearing a costume.
Why nothing adds up (and why that is normal)
Four systems will give you four numbers for the same month. Before concluding that the tracking is broken, understand that each is answering a different question.
| System | What it counts | The rule that trips people up |
|---|---|---|
| Google Ads | Conversions it can tie back to an ad click | It credits the conversion to the date of the click, not the date of the sale. A sale today from a click three weeks ago is reported three weeks ago — which is why last month’s numbers keep changing after month-end. |
| GA4 | Events on your site, from tagged sessions | It records events when they happen and applies its own attribution model, which affects event-scoped dimensions only. Google states that “user- and session-scoped traffic dimensions, such as Session source or First user medium, are unaffected by changes to the reporting attribution model” — so two GA4 reports can legitimately disagree. |
| Meta | Conversions inside its own attribution windows | The rules changed twice in 2026: longer view-through windows were removed in January, and from 3 March a click-through requires an actual link click, with likes, saves, comments and video engagement moved to a new one-day “engage-through” bucket. |
| Your CRM | People, and what happened to them | It only knows what someone typed in. It is also the only system that knows whether the lead was any good. |
Two consequences worth writing into your own reporting. First, never add platform conversions together. Meta and Google will both claim the same sale, so a “total conversions” figure summed across platforms is a number that does not exist. Second, any comparison spanning early 2026 crosses a definition change — if Meta click-through conversions fell in March, the first question is not “what went wrong” but “how much of this is reclassification”.
The deeper mechanics are covered in the performance marketing guide, and the specific case of social is in measuring social media ROI.
The Singapore-specific distortions
GST is sitting inside two of your numbers
If your checkout passes the gross order value to the pixel, the revenue in your ROAS includes the 9% GST you will remit to IRAS. On the cost side, Google Ads has charged 9% GST on Singapore-billed accounts since 1 January 2024, while under Singapore’s Overseas Vendor Registration rules a GST-registered business that supplies its GST number to a registered overseas supplier should not be charged GST on those digital services. Two businesses running identical campaigns can therefore have genuinely different true costs depending on their GST status — and neither difference appears in the platform report. Full working in ROAS vs ROI.
Consent gaps make your platform numbers look worse than reality
Under the PDPA, consent is required for cookies used for advertising and targeting purposes, and the legitimate interests exception does not cover direct marketing. Visitors who decline are invisible to your pixel but can still buy. That widens the gap between platform-reported conversions and CRM reality in a way that is a measurement artefact, not lost business. A good report says so; a weak one lets you assume the campaign underperformed.
Seasonality here is unusually lumpy
Chinese New Year, the mid-year sales period, Ramadan and Hari Raya, the 11.11 and 12.12 platform events and the year-end school holidays all move Singapore demand hard, on different calendars for different sectors. A month-on-month comparison that ignores which festivals fell where is close to meaningless. Ask for the year-on-year view alongside it, with the caveat about 2026’s definition changes attached.
The eight questions that make a reporting call useful
- “What was the target, and did we hit it?” If there was no target, that is the finding. Set one before the next call.
- “How many of those leads were qualified, by our written definition?” The single highest-value question in the meeting.
- “What did a qualified enquiry cost us, all in?” Media, fees and tools — not just media.
- “Which of these numbers comes from which system?” Forces the sourcing into the open, and usually explains half the confusion.
- “What did we test this month, and what did it tell us?” No tests means no learning, whatever the arrows did.
- “What did not work, and what did that cost?” Ask it every month until it stops feeling awkward.
- “What would you do if this were your money and the budget were 30% smaller?” Unusually revealing. The answer names the spend nobody can defend.
- “What are the three specific things happening next month, and who owns each?” Vague forward plans are how a flat quarter becomes a flat year.
If your report only ever shows the top bar, you are being invited to judge the marketing on the widest, most flattering number available. Ask for the bottom bar. If nobody can produce it, that is the first project — not a better ad.
Red flags, and what an honest answer sounds like
| Red flag | What it usually means | What a straight answer sounds like |
|---|---|---|
| Page one leads with reach or impressions | The outcome numbers were not flattering this month | “Leads were 18% below target. Here is why, and here is what we are changing.” |
| Percentages with no absolute numbers | The base is small. “Up 200%” can mean two leads became six. | “Enquiries went from 11 to 19, so up 73% off a small base — treat it as directional.” |
| Conversions summed across platforms | Nobody has reconciled the systems | “Meta claims 40, Google claims 25, the CRM logged 47 in total. Here is our best reconciliation.” |
| Retargeting is the star of the report | Credit is being taken for people who were already going to buy | “Retargeting shows the best ROAS but is the least incremental. Here is how we would test it.” |
| Nothing failed this month | Either nothing is being tested, or results are being filtered | “The new lead form lost. It cost about S$900 and two weeks. Here is what we learned.” |
| The same recommendations every month | The report is a template, not an analysis | Next month’s actions differ from last month’s and reference what actually happened. |
| Claims that ad spend or the retainer is grant-claimable | A misunderstanding, or worse | “Ad spend and retainers are generally not claimable. PSG covers pre-approved solutions, and you apply yourself.” |
What should change between month one and month six
A reporting relationship that is working matures visibly. In the first month or two the report is necessarily about setup and baselines — tracking installed, definitions agreed, benchmarks established, and an honest acknowledgement that the numbers are not yet trustworthy. By month three you should be seeing outcomes against a target and the first genuine tests. By month six the report should be shorter, not longer: metrics that never changed a decision should have been dropped, and the conversation should have narrowed to the few things that actually move.
If your month-six report looks exactly like your month-one report with different numbers, nothing has been learned. That is the clearest signal available, and spotting it requires no technical knowledge at all.
A worked example of a good summary page
To make this concrete, here is what the first page of a useful monthly report looks like for a Singapore services business, in about eighty words:
“Target was 40 qualified enquiries at under S$150 each. We delivered 34 at S$168, so we missed on both. The shortfall is entirely in the Meta lead-form campaign, where cost per qualified enquiry rose from S$121 to S$214 after the audience saturated — frequency reached 4.6. Search held steady at S$139. We paused two Meta ad sets on the 18th and shifted S$1,800 to search. Next month we are testing two new creative angles and a landing page for the higher-value service line.”
Note what is present: a target, a miss stated plainly, the specific cause, the diagnostic number that supports it, the action already taken, and the plan. Note what is absent: impressions, reach, and any adjective doing work that a number should be doing.
Why we report the way we do
Our approach is built around a report an owner can read without a translator: outcomes first, efficiency second, the plain-English “why” third, an honest account of what failed, and a specific plan with owners. We name which system each number came from, and we do not add platform conversions together. When a number moved because a platform changed its definition, we say so rather than taking credit or blame for it.
The results we have been able to publish are in our case studies, and the reporting discipline itself is part of our performance marketing service.
Where to go next
- Performance marketing in Singapore — the hub for this cluster.
- ROAS vs ROI — the two numbers most often confused in a report.
- Measuring social media ROI — where the reconciliation problem is worst.
- Google Ads costs and Meta Ads costs in Singapore — for sanity-checking the spend side.
- When Google Ads are not converting — the diagnostic path when the report shows traffic but no outcomes.
The short version
A marketing report is a decision document. If it does not tell you what happened against a target, why, what failed, and what happens next, it is a status update in a nicer font. You do not need to learn attribution theory to hold one to account — you need a written definition of a lead, one system nominated as the scoreboard, and the willingness to keep asking what did not work until asking it stops feeling rude.
Tired of reports you cannot decode? Talk to us and we will show you what your current reporting is leaving out, or read how we run performance marketing.
Frequently asked questions
What metrics actually matter in a marketing report?
Qualified enquiries or sales against a target, the all-in cost per qualified outcome and its trend, a plain-English explanation of why it moved, an honest account of what failed, and specific next actions with owners. Impressions, reach and clicks belong in the diagnostics section, not on page one.
What are vanity metrics?
Numbers that look impressive but do not change a decision — impressions, reach, follower counts, and clicks quoted in isolation. They are not useless as diagnostics; they are misleading as headlines, because they are the easiest metrics to improve without improving the business.
Why do Google, Meta and my CRM show different numbers?
Google Ads credits conversions to the date of the ad click rather than the date of the sale. GA4 records events when they happen and applies its own attribution model, and its user- and session-scoped dimensions are unaffected by the reporting model setting. Meta counts within its own windows, which changed in January and March 2026. Your CRM counts people. Never add them together.
How do I know if my agency’s reporting is trustworthy?
It leads with outcomes against a target, names the source system for each number, includes something that failed every month, does not sum conversions across platforms, and its recommendations differ from last month’s. A report that gets shorter over time is usually a good sign.
Should my report show ROAS or cost per lead?
Whichever maps to how you make money. E-commerce should see ROAS alongside a break-even figure derived from gross margin. Service businesses should see cost per qualified enquiry, ideally with close rate and average job value so the number can be converted into profit. Seeing both, with neither defined, is a warning sign.
How often should I get a marketing report?
Monthly suits most Singapore SMEs. Weekly reporting on small budgets produces noise that invites over-reacting, because a single week rarely contains enough conversions to be meaningful. Ask instead for monthly reporting plus a message when something breaks or a test concludes.
Related measurement guides
- Conversion tracking in Singapore: GA4, Google Ads and Meta
- Attribution models explained: which channel actually made the sale
Related measurement guides
Related measurement guides
- Vanity metrics in Singapore marketing — the three-tier test for deciding which numbers deserve the front page of this report.
- PDPA-compliant marketing tracking — what Singapore law requires of the pixels and tags producing these numbers.
